|
2
|
1 <?xml version="1.0" encoding="UTF-8"?>
|
|
|
2 <tool id="edger_dge" name="edgeR Differential GeneExpression Analysis">
|
|
|
3 <description>RNA-Seq expression analysis using edgeR (R package)</description>
|
|
|
4
|
|
22
|
5 <command>
|
|
2
|
6 <!--
|
|
|
7 The following script is written in the "Cheetah" language:
|
|
|
8 http://www.cheetahtemplate.org/docs/users_guide_html_multipage/contents.html
|
|
|
9 -->
|
|
|
10
|
|
15
|
11 R --vanilla --slave -f $R_script '--args
|
|
2
|
12 $design_matrix
|
|
|
13 $contrast
|
|
|
14
|
|
|
15 $output_count_edgeR
|
|
|
16 $output_cpm
|
|
|
17 output_FPXM
|
|
|
18 $output_raw_counts
|
|
|
19
|
|
|
20 $qc
|
|
|
21 $output_MDSplot
|
|
|
22 $output_BCVplot
|
|
|
23 $output_MAplot
|
|
|
24 smearPlot '
|
|
21
|
25 > $output_R
|
|
22
|
26 2> stderr.txt ;
|
|
|
27 grep -v 'Calculating library sizes from column' stderr.txt 1>&2
|
|
|
28
|
|
2
|
29 </command>
|
|
|
30
|
|
|
31 <inputs>
|
|
|
32 <param name="design_matrix" type="data" format="tabular" help="Design matrix" />
|
|
|
33
|
|
|
34 <param name="contrast" type="text" label="Contrast (biological question)" help="e.g. 'tumor-normal' or '(G1+G2)/2-G3' using the factors chosen in the design matrix. Read the 'makeContrasts' manual from Limma package for more info." />
|
|
|
35
|
|
|
36 <param name="qc" type="select" label="Quality control reports">
|
|
18
|
37 <option value="true" selected="true">Yes</option>
|
|
|
38 <option value="false">No</option>
|
|
2
|
39 </param>
|
|
|
40
|
|
|
41 <param name="debug" type="select" label="R Debug output">
|
|
18
|
42 <option value="true"> Yes</option>
|
|
|
43 <option value="false" selected="true">No</option>
|
|
2
|
44 </param>
|
|
|
45 </inputs>
|
|
|
46
|
|
4
|
47 <configfiles>
|
|
|
48 <configfile name="R_script">
|
|
16
|
49 library(edgeR,quietly=TRUE) ## enable quietly to avoid unnecessaity stderr dumping for loading limma
|
|
4
|
50
|
|
|
51 ## Fetch commandline arguments
|
|
6
|
52 args <- commandArgs(trailingOnly = TRUE)
|
|
4
|
53 designmatrix = args[1]
|
|
|
54 contrast = args[2]
|
|
|
55
|
|
|
56 output_1 = args[3]
|
|
|
57 output_2 = args[4]
|
|
15
|
58 output_3 = args[5] ##FPKM file - yet to be implemented
|
|
4
|
59 output_4 = args[6]
|
|
|
60
|
|
|
61 QC = nchar(args[7]) > 0
|
|
|
62
|
|
|
63 output_5 = args[8]
|
|
|
64 output_6 = args[9]
|
|
|
65 output_7 = args[10]
|
|
|
66
|
|
|
67 output_8 = args[11]
|
|
|
68
|
|
|
69
|
|
|
70 library(edgeR)
|
|
6
|
71 raw_data <- read.delim(designmatrix,header=T,stringsAsFactors=T)
|
|
4
|
72
|
|
|
73 ## Obtain read-counts
|
|
12
|
74
|
|
|
75 header = read.delim(as.character(raw_data[1,1]),header=F,stringsAsFactors=F,row.names=1,nrows=1)
|
|
|
76 has_header = (class(header[1,1]) == "character")
|
|
|
77
|
|
13
|
78 read_counts = read.delim(as.character(raw_data[1,1]),header=has_header,stringsAsFactors=F,row.names=1)[1]
|
|
12
|
79
|
|
4
|
80 for(i in 2:length(raw_data[,1])) {
|
|
15
|
81 write("parsing counts from:",stdout())
|
|
|
82 write(raw_data[i,1],stdout())
|
|
11
|
83
|
|
|
84 header = read.delim(as.character(raw_data[i,1]),header=F,stringsAsFactors=F,row.names=1,nrows=1)
|
|
|
85 has_header = (class(header[1,1]) == "character")
|
|
13
|
86 table = read.delim(as.character(raw_data[i,1]),header=has_header,stringsAsFactors=F,row.names=1)[1]
|
|
11
|
87
|
|
13
|
88 read_counts = cbind(read_counts,table)
|
|
4
|
89 }
|
|
|
90
|
|
13
|
91 colnames(read_counts) = as.character(raw_data[,2])
|
|
|
92
|
|
4
|
93
|
|
|
94
|
|
|
95 ## Filter for HTSeq predifined counts:
|
|
|
96 exclude_HTSeq = c("no_feature","ambiguous","too_low_aQual","not_aligned","alignment_not_unique")
|
|
|
97 exclude_DEXSeq = c("_ambiguous","_empty","_lowaqual","_notaligned")
|
|
|
98
|
|
|
99 exclude = match(c(exclude_HTSeq, exclude_DEXSeq),rownames(read_counts))
|
|
|
100 exclude = exclude[is.na(exclude)==0]
|
|
|
101 if(length(exclude) != 0) {
|
|
|
102 read_counts = read_counts[-exclude,]
|
|
|
103 }
|
|
|
104
|
|
|
105
|
|
|
106
|
|
13
|
107 empty_samples = apply(read_counts,2,function(x) sum(x) == 0)
|
|
|
108 if(sum(empty_samples) > 0) {
|
|
|
109 write(paste("There are ",sum(empty_samples)," empty samples found:",sep=""),stderr())
|
|
|
110 write(colnames(read_counts)[empty_samples],stderr())
|
|
14
|
111 } else {
|
|
13
|
112 dge = DGEList(counts=read_counts,genes=rownames(read_counts))
|
|
4
|
113
|
|
13
|
114 design_tmp <- raw_data[3:length(raw_data)]
|
|
|
115 rownames(design_tmp) <- colnames(dge)
|
|
|
116 formula = paste(c("~0",colnames(design_tmp)),collapse = " + ")
|
|
|
117 design <- model.matrix(as.formula(formula),design_tmp)
|
|
4
|
118
|
|
13
|
119 prefixes = colnames(design_tmp)[attr(design,"assign")]
|
|
|
120 avoid = nchar(prefixes) == nchar(colnames(design))
|
|
|
121 replacements = substr(colnames(design),nchar(prefixes)+1,nchar(colnames(design)))
|
|
|
122 replacements[avoid] = colnames(design)[avoid]
|
|
|
123 colnames(design) = replacements
|
|
4
|
124
|
|
|
125
|
|
|
126
|
|
15
|
127 write("Calculating normalization factors...",stdout())
|
|
13
|
128 dge = calcNormFactors(dge)
|
|
15
|
129 write("Estimating common dispersion...",stdout())
|
|
13
|
130 dge = estimateGLMCommonDisp(dge,design)
|
|
15
|
131 write("Estimating trended dispersion...",stdout())
|
|
13
|
132 dge = estimateGLMTrendedDisp(dge,design)
|
|
15
|
133 write("Estimating tagwise dispersion...",stdout())
|
|
13
|
134 dge = estimateGLMTagwiseDisp(dge,design)
|
|
4
|
135
|
|
|
136
|
|
|
137
|
|
|
138
|
|
15
|
139 if(QC == TRUE) {
|
|
|
140 write("Creating QC plots...",stdout())
|
|
13
|
141 #### MDS Plot
|
|
|
142 pdf(output_5)
|
|
|
143 plotMDS(dge, main="edgeR MDS Plot")
|
|
|
144 dev.off()
|
|
|
145 #### Biological coefficient of variation plot
|
|
|
146 pdf(output_6)
|
|
|
147 plotBCV(dge, cex=0.4, main="edgeR: Biological coefficient of variation (BCV) vs abundance")
|
|
|
148 dev.off()
|
|
|
149 }
|
|
4
|
150
|
|
|
151
|
|
|
152
|
|
15
|
153 write("Fitting GLM...",stdout())
|
|
13
|
154 fit = glmFit(dge,design)
|
|
4
|
155
|
|
15
|
156 write(paste("Performing likelihood ratio test: ",contrast,sep=""),stdout())
|
|
13
|
157 cont <- c(contrast)
|
|
|
158 cont <- makeContrasts(contrasts=cont, levels=design)
|
|
4
|
159
|
|
13
|
160 lrt <- glmLRT(fit, contrast=cont[,1])
|
|
15
|
161 write(paste("Exporting to file: ",output_1,sep=""),stdout())
|
|
13
|
162 write.table(file=output_1,topTags(lrt,n=nrow(read_counts))\$table,sep="\t",row.names=T)
|
|
|
163 write.table(file=output_2,cpm(dge,normalized.lib.sizes=TRUE),sep="\t")
|
|
|
164 ## todo EXPORT FPKM
|
|
|
165 write.table(file=output_4,dge\$counts,sep="\t")
|
|
15
|
166
|
|
|
167
|
|
|
168
|
|
|
169 if(QC == TRUE) {
|
|
|
170 write("Creating MA plots...",stdout())
|
|
13
|
171
|
|
|
172 etable <- topTags(lrt, n=nrow(dge))\$table
|
|
|
173 etable <- etable[order(etable\$FDR), ]
|
|
|
174 pdf(output_7)
|
|
|
175 with(etable, plot(logCPM, logFC, pch=20, main="edgeR: Fold change vs abundance"))
|
|
|
176 with(subset(etable, FDR<0.05), points(logCPM, logFC, pch=20, col="red"))
|
|
|
177 abline(h=c(-1,1), col="blue")
|
|
|
178 dev.off()
|
|
|
179 }
|
|
15
|
180 write("Done!",stdout())
|
|
4
|
181 }
|
|
|
182 </configfile>
|
|
|
183 </configfiles>
|
|
|
184
|
|
2
|
185 <outputs>
|
|
|
186 <data format="tabular" name="output_count_edgeR" label="edgeR DGE on ${design_matrix.hid}: ${design_matrix.name} - table" />
|
|
|
187 <data format="tabular" name="output_cpm" label="edgeR DGE on ${design_matrix.hid}: ${design_matrix.name} - CPM" />
|
|
|
188 <data format="tabular" name="output_raw_counts" label="edgeR DGE on ${design_matrix.hid}: ${design_matrix.name} - raw counts" />
|
|
|
189
|
|
|
190 <data format="txt" name="output_R" label="edgeR DGE on ${design_matrix.hid}: ${design_matrix.name} - R output" >
|
|
|
191 <filter>(debug == "true")</filter>
|
|
|
192 </data>
|
|
|
193
|
|
|
194 <data format="pdf" name="output_MDSplot" label="edgeR DGE on ${design_matrix.hid}: ${design_matrix.name} - MDS-plot">
|
|
|
195 <filter>(qc == "true")</filter>
|
|
|
196 </data>
|
|
|
197
|
|
|
198 <data format="pdf" name="output_BCVplot" label="edgeR DGE on ${design_matrix.hid}: ${design_matrix.name} - BCV-plot">
|
|
|
199 <filter>(qc == "true")</filter>
|
|
|
200 </data>
|
|
|
201
|
|
|
202 <data format="pdf" name="output_MAplot" label="edgeR DGE on ${design_matrix.hid}: ${design_matrix.name} - MA-plot">
|
|
|
203 <filter>(qc == "true")</filter>
|
|
|
204 </data>
|
|
|
205 </outputs>
|
|
|
206
|
|
|
207 <help>
|
|
|
208 input: Design matrix using "create Design matrix" tool
|
|
|
209 input: contrast
|
|
|
210 </help>
|
|
|
211 </tool>
|
